ANA MARIA CARVALHO 642

VARIATION AND DIFFUSION OF URUGUAYAN PORTUGUESE IN A BILINGUAL BORDER TOWN

Ana Maria Carvalho1 University of California at Berkeley, USA

In Northern , along the Brazilian border, Portuguese was the only language spoken by the Portuguese and the Brazilian settlers. It was only at the end of the last century that Spanish was imposed upon the Portuguese speaking communities through state educational policies and language planning. Spanish was then brought to the border areas as a prestigious language, and rural Portuguese was maintained as the vernacular. This diglossic situation continues into the present; generally, Portuguese is employed in familiar circumstances, whereas Spanish is primarily the language of public life. However, as Elizaincín points out (1978: 304), while the entire population follows diglossic rules, patterns of language choice vary within this model according to the individual’s social status. Thus, while the working class uses Spanish only in situations where Spanish is required (work, education, church), and Uruguayan Portuguese in all in-group activities, the middle-class prefers to use Spanish most of the time. The greater use of the ‘official language’ by the middle class separates them from the lower social classes, and, as noted by Behares (1984), marks their status as a group who identifies with the national culture rather than with that of the border. Meanwhile, Uruguayan Portuguese works as a unifying factor for the working-class, signing solidarity and self identification with the local culture. My analysis indicates that besides language choice, internal variation within Uruguayan Portuguese also reflects this social stratification. Pedro Rona studied ‘fronterizo’, a variety spoken by monolinguals, which he defined as “una mezcla de portugués y español, pero que no es ni portugués ni español, y resulta con frecuencia ininteligible tanto para los brasileños como para los uruguayos”2 (1965: 7). Fritz Hensey, in 1972, adopted the term ‘fronterizo’ proposed by Rona to designate a “set of which are spoken in mostly rural sectors of Northern Uruguay often as a sole language”. However, he chose to study what he defined as “the Portuguese spoken by bilinguals in the town of Rivera”. Elizaincín rejected the term ‘fronterizo’ and proposed instead DPU (Dialectos Portugueses del Uruguay), which he defined as “formas mixtas, de

1 1052 9th St. Albany, CA. USA, 94710; E-mail: [email protected]

2 “A mixture of Spanish and Portuguese, but which is neither one, and it is not understandable by Brazilians and ”. ACTAS DO I SIMPOSIO INTERNACIONAL SOBRE O BILINGÜISMO 643 base preponderantemente portuguesa las que, sin embargo, evidencian fuerte influencia del español”3 (1987: 14). According to Elizaincín, DPU varieties are spoken by Rivera’s lower social classes, usually as their sole language. The general question which motivated my study was whether we could actually separate Hensey’s Uruguayan Portuguese of bilinguals, from Rona’s Fronterizo and Elizaincín’s DPU. Given the social and historical characteristics of Rivera, the border town where I carried out my research, I thought that perhaps the abstraction of a linguistic continuum, which would consist of a set of possible choices ranging from rural Uruguayan Portuguese to standard Portuguese, could bridge these varieties. This continuum, together with border Spanish and its own set of variation, would then better describe the linguistic repertoire of this bilingual community. Detecting these linguistic continua and matching their sections to sociological and stylistic factors, in addition to revealing patterns of choice between Spanish and Portuguese, is the aim of my dissertation entitled “The Social distribution of Spanish and Portuguese dialects in the bilingual town of Rivera, Uruguay”. Today, I would like to present some preliminary results and theoretical assumptions that form the basis of my analysis of the Uruguayan Portuguese spoken in Rivera. More specifically, I will argue that recent social changes such as urbanization and greater exposure to the standard model have affected the local variety, which is now undergoing a change from the highly focused rural dialect to a less stigmatized urban version. This process is analogous to what Bortoni-Ricardo (1985) encountered in the urbanization of rural dialects in Brazil and defined as dialect diffusion, a concept introduced by Le Page (1975). The transition from dialect focusing, which consists of higher incidence of local, stigmatized variants, to dialect diffusion, enables some speakers to sound more local or less, depending on the choice of lexical, morphosyntactic, and phonological variants. The quantification of two phonological variables will allow us to identify the social groups leading the diffusion and make assumptions about their motivations. I will conclude that subgroups of the sample population show different degrees of identification with the border culture, and the use of linguistic norms associated with the border reflect such differences. These preliminary conclusions are based on results of field work in Rivera. Rivera is a town on the northeast side of the border. This is an open border, and movement of cars and people between Rivera and Sant’anna do Livramento, Rivera’s Brazilian twin city, is uncontrolled. For this paper, 36 of the interviews in Portuguese were analyzed. The

3 “Mixed forms, with a predominantly Portuguese base, which, however, show high degree of Spanish influence”. ANA MARIA CARVALHO 644 interviewees consist of an equal number of men and women and represent three socio- economic classes and age groups. Although I am currently inserting my data in the statistics package VARBRUL and the final results are not ready yet, for this paper I have calculated the average mean for each group in order to illustrate what the statistical results will show. In order to understand Rivera’s current linguistic profile, it is important to refer to its recent history. The last quarter century has brought significant changes to Rivera. Rivera began as a small town, in which a large part of the population was monolingual in Uruguayan Portuguese with close ties to the countryside and limited contacts with the Uruguayan South and with urban Brazil. It has since become an important regional center attracting tourists and shoppers from both sides of the border. As Hensey reports (1982: 18), Portuguese monolingualism was replaced by bilingualism due to easier access to the education system in which Spanish is taught, in addition to the opening of institutes and businesses that brought to the border professionals from the Spanish-speaking South. Moreover, the dialects spoken in urban Brazil are also increasingly present due to the establishment of several duty-free shops in the 80’s, which has made Rivera a potential tourist and shopping center for Southern Brazilians. In addition to greater interaction with outsiders, perhaps the most powerful presence of Brazilian urban dialects in this community comes through television. Brazilian tv programs and soap operas are part of the Riverenses’ daily routine. Thus, urbanization has increased this community’s contact with normative agencies of both languages. Consequently, the variety of Portuguese traditionally spoken in these communities and characterized by Rona (1965) and Elizaincín (1975, 1979, 1992) as rural and highly influenced by Spanish, has been pulled in the direction of the more prestigious urban Brazilian Portuguese model. This tendency can be seen through the local repertoire’s incorporation of new variants that have been borrowed from Brazilian Portuguese at the expense of local variants, through dialect diffusion. This assumption is based on the high prestige assigned to the monolingual varieties of Brazilian Portuguese in this community. Uruguayan Portuguese, on the other hand, is overtly stigmatized for several reasons. It is rural; it is non-official; and it is heavily influenced by Spanish, which contradicts the popular notion of linguistic purity. Consequently, Riverenses are linguistically insecure, and this insecurity is exacerbated by the constant presence of the standard model. Their negative attitudes towards their Portuguese, the great fluctuation across styles, and their constant self-correction characterize what Labov defines as linguistic-self hatred (1966). In what follows I reproduce some typical comments. One informant, a primary school teacher, said: ACTAS DO I SIMPOSIO INTERNACIONAL SOBRE O BILINGÜISMO 645

Me dicen que tenía que tener orgullo de mi lengua materna, pero es una lengua materna tan fea! I was told I should be proud of my mother language, but it is such an ugly mother language!

Maybe to compensate this “ugliness”, another said:

Nós tratemo de imitar o português do Brasil. We tried to imitate Brazilian Portuguese.

Another informant, proving how influential the television can be, said:

Gosto do jeito que o pessoal fala nas novela porque é brasileiro em si, entonces eu trato de imitar. Nós aqui somo rompe-idioma. I like the way the people talk in the soap operas because it is real Brazilian. So I try to imitate it. Here, we are language breakers.

Moreover, the dialect spoken on the other side of the border is seen as similar to the one on tv, as shown in the following comment by a teenager who works at an expensive bakery, which attracts many tourists: Os brasilero que vem na padaria falam como os da televisão, não como nós aqui em Rivera e na campanha, que é mesclado. The Brazilians that come here into the bakery speak like the people on tv, and not like us here in Rivera or in the countryside, which is mixed.

Thus, contrary to Elizaincín’s belief that Uruguayan Portuguese standard speakers lack standard models (1992: 230), the presence of Brazilian Portuguese in this community provides the linguistic model for Rivera speakers. That is, for those who want it, since the standardizing force of Brazilian Portuguese is assimilated in varying degrees by the different social groups. This results in the coexistence of local and borrowed forms which constitutes the current linguistic variation. Depending on the speaker’s social characteristics, this variation will show either focusing around the genuine local variety or diffusion towards the Brazilian model. This variation between local and borrowed forms can be represented by a continuum, as I proposed earlier (Figure 1). On the left end, we have the forms that characterize genuine rural border Uruguayan Portuguese, and on the right, we have forms that characterize the standard counterpart. On each of these extremes, the features are categorical and create a sharp distinction between these two varieties of Portuguese. We may think of Rivera’s speakers as being in between these two extremes, and their speech as tending towards the left or the right depending on social and stylistic characteristics, which motivate them to diffuse or to focus.

ANA MARIA CARVALHO 646

Figure 1: Dialectal continuum of Rural Uruguayan Portuguese (RUP) and Urban Brazilian Portuguese (UBP)

RUP------UBP Lexicon: Rural forms No rural forms Spanish borrowings No Spanish borrowings Border neologisms No border neologisms Morphos: Rural features No rural features Spanish interference No Spanish interference Phonol: Rural forms No rural forms Spanish interference No Spanish interference

focusing Õ Ö diffusing

In all categories of grammar, lexicon, morphosyntax, and phonology, Uruguayan Portuguese differs from the standard Brazilian variety mainly in three respects. First, it contains rural features which are no longer present in Brazilian cities. Second, Spanish interference is common, which contributes to the popular characterization of it as ‘portuñol’. And finally, it presents hybrid forms that cannot be found in either language but are peculiar to this variety. The presence of borrowings, rural forms, and neologisms characterize Uruguayan Portuguese, and although its distribution is outside the scope of this paper, their use should be considered polarized examples of a focused dialect. Their absence, on the other hand, is due to the careful substitution of the equivalents in Standard Portuguese, indicating dialect diffusion. The two phonological variables chosen to illustrate dialect diffusion and focusing are the vocalization of the palatal liquid //, a rural form, and the maintenance of the dental stops /d/ and /t/ followed by /i/, a tendency kept probably due to Spanish interference. Because no linguistic conditioning was found, I will present only a summary of their relation to social and stylistic variation. The vocalization of the palatal liquid, that is, the pronunciation of the lateral as a front glide, as in /muj/ instead of the standard /mur/ has been classified by Bortoni- Ricardo (1985: 174) as a rural stereotype in Brazil and was ranked the most stigmatized variant of non-standard Portuguese by Brian Head (1981: 164). The vocalization is never present in Brazilian television programs or soap operas. In Rivera Portuguese, this vocalization was detected by Rona (1965: 23) as categorical, which led him to propose the glide as the only possible phoneme. My data reveal that this is a stratified sociolinguistic feature and a marker of local, focused Portuguese. In contrast, the standard palatal realization is linked to urban Brazilian Portuguese, and its use is a sign of dialect diffusion. The ACTAS DO I SIMPOSIO INTERNACIONAL SOBRE O BILINGÜISMO 647 vocalization is, in fact, a linguistic stereotype of Rivera Portuguese since all my respondents recognize it as being typical of the local dialect. Indeed, one informant, when showed the picture of a spoon during the picture naming section, said: Em brasilero, colher, na fronteira, coié. In Brazilian, colher, on the border, coié.

The realization of this variable as either the local glide or the “Brazilian” lateral is, thus, a strong indicator of dialect diffuseness or focusing. Because vocalization is reported to have been categorical in border Portuguese, and the pronunciation of the liquid consonant an innovation, I dealt with vocalization as the underlying form, and consonant as the marked one. Let us examine the use of the consonantal phoneme according to social groups’ average in the speech sample:

Table 1 - Rate of use of the palatal liquid [] for orthographic lh in intervocalic positions according to social class

Conversational style Picture naming

Frequency % Frequency % working 369/75 24.0 76/65 83.83 low-middle 345/161 46.45 73/65 89.29 middle-middle 175/141 75.62 83/78 94.05

Table 2 - Rate of use of the palatal liquid [] for orthographic lh in intervocalic positions according to age

Conversational style Picture naming

Frequency % Frequency % (50-70) 321/130 39.86 76/63 81.05 (29-49) 274/109 47.83 76/70 92.46 (15-28) 294/138 58.38 80/75 93.65

Table 3 - Rate of use of the palatal liquid [] for orthographic lh in intervocalic positions according to gender

Conversational style Picture naming

Frequency % Frequency % Male 462/136 34.87 115/96 82.21 Female 427/241 62.51 114/112 95.90

Total St. Deviation Total St. Deviation 0.37 0.16

ANA MARIA CARVALHO 648

As it can be seen, the sharp difference of percentages between the conversational style and the picture-naming sections proves that this is an important stylistic marker: informants across groups show higher scores in the formal style, picture naming. Moreover, the prestigious // is much less frequent in the speech of the lowest social class, the oldest group, and men, all of whom speak a more focused dialect. Meanwhile, the speech of the middle- middle class, the young, and women show that these are the groups whose Portuguese is more diffused towards the prestigious model. The second diagnostic variable is the palatalization of dental stops. In most dialects of Brazilian Portuguese, the dental stops have realizations whenever they occur preceding the high front vowel /i/. Therefore, the word ‘dia’ is most likely to be pronounced as [día] instead, and ‘tia’ is pronounced [tía]. Leda Bisol (1992) studies the application of the palatalization rule in the Brazilian South and observes that this phenomenon has been expanding in the area. This is the result of geographical diffusion which probably started in Rio de Janeiro. In comparing my own data with previous studies of Uruguayan Portuguese, it becomes clear that the application of palatalization is being slowly incorporated in this variety. Rona (1965: 40) observes that, in Rivera, the dental realization was nearly categorical. Hensey (1972: 60) later proves this by demonstrating that Riverenses palatalized much less than border Brazilians. For Riverenses, palatalization is a linguistic stereotype of the urban monolingual Brazilian accent. In fact, a male speaker, whose speech is focused, when asked about the difference between Uruguayan Portuguese and the variety spoken on TV answered:

No português da televisão há muito [t]. Dizem “para [ti]” “[tira]. Aqui é [ti], [tira]. Yo hablo portuñol, não falo [ti]. In the Portuguese spoken on television there are many [ti]. They say [para ti], [tira]. I speak portunol, I never say [ti].

This statement shows that this variable is a marker of the local dialect, and that palatalization is an indicator of dialect diffusion. It shows that the speakers are usually be able to differentiate the local dialect from the standard one according to the application, or not, of the palatalization rule. We may conclude from this that, besides being aware of this difference, speakers are also able to make the decision on which variant to use. Moreover, they base their decision on the variant’s social value and whether or not they choose to sound like a local, a speaker of “portuñol”.

ACTAS DO I SIMPOSIO INTERNACIONAL SOBRE O BILINGÜISMO 649

Table 4: Rate of palatalization of [di] and [ti] according to social class:

Conversational Style Picture-naming

Frequency % Frequency % Working 372/54 12.28 49/15 25.69 Low-middle 421/109 19.12 49/20 36.98 Middle-middle 398/220 39.33 44/23 44.58

Table 5: Rate of palatalization of [di] and [ti] according to age:

Conversational Style Picture-naming

Frequency % Frequency % (50-70) 277/8 4.71 42/3 6.53 (29-49) 344/42 10.10 44/10 20.97 (15-28) 570/333 55.92 56/45 79.76

Table 6: Rate of palatalization of [di] and [ti] according to gender:

Conversational Style Picture-naming

Frequency % Frequency % Male 539/44 18.06 70/27 34.84 Female 652/239 29.09 72/31 36.67

Total St. Deviation Total St. Deviation 0.31 0.40

Tables 4, 5 and 6 show the percentage of palatalization in different social groups. It is worth noting that this variable is not subjected to a drastic stylistic difference as was true for the liquid. Unlike the distribution of the liquid, palatalization seems to be a variable related predominantly to age. The significantly more frequent occurrences of palatalization in the younger groups suggest that this is an innovation in the dialect, which has generally followed the conservative tendency to produce dentals as stops, a tendency maintained in the speech of the elderly. Following a diffusion that originated in central urban Brazil, the prestige of this variant derives from its social and geographic origin. It is a typical case of language change caused by the spread of prestige patterns of urban capitals, which, in this case, crosses national borders. The pronunciation of this variable as a dental stop is not felt to be stigmatized, as is the case of the vocalization of the liquid. However, it is definitely a feature of the local dialect, and its substitution for the palatalized version is a way by which Rivera youth attempt to align themselves with a group that is different from their parents. ANA MARIA CARVALHO 650

One more word should be said about this variable: the application of palatalization may pose a difficulty of comprehension for speakers of Spanish ([tia], for example, is hardly recognized as “tia”). Therefore, the incorporation of the palatalization rule in the Rivera speech runs against the universal constraint on linguistic change that, in contact situations, mergers expand at the expense of distinctions, as argued by Herzog (1965: 211), and further discussed by Labov (1972: 300). In the case of Rivera, rather than neutralizing differences to enhance mutual comprehension, we found differences being incorporated in order to mark social status. In conclusion, a comparison of our results with previous studies demonstrates a higher incidence of phonological variants associated with Brazilian Portuguese. We interpret this as the result of diffusion affecting the grammar of the community. This diffusion is demonstrated here by Riverenses’ avoidance of phonological stereotypes perceived as their “way of talking” and is present in the speech of the groups, who, we have hypothesized, lack identification with the border. Their social characteristics, such as high social status, young age, and female gender, are amplifiers (a term coined by Bailey et al., 1993) for the diffusion. Meanwhile, the elderly, the men, and the working-class’ local pride, solidarity, and identification with the border, work as barriers, impeding the diffusion of these linguistic innovations, and perpetuating the border linguistic heritage. As proposed by Le Page and discussed in Bortoni- Ricardo, “speakers create their rules so as to resemble as closely as possible the members of the group with which they wish to identify” (1985: 93). In Rivera, the groups that identify with the border culture speak a more focused variety of Portuguese, and more often choose it over Spanish. On the other hand, the groups that would prefer to be identified with the mainstream Uruguayan culture speak more Spanish than Portuguese, along with a more diffused Uruguayan Portuguese, as a signal of identification also with the urban, mainstream Brazilian culture. The innovations brought to the border dialect through diffusion once again prove Thompson and Kaufman correct in their assertion that, “it is the sociolinguistic history of the speakers, and not the structure of their language, that is the primary determinant of the linguistic outcome of language contact” (1991: 35).

References Bailey, G., T. Wikle, J. Rillery, L. Sand (1993), “Some patterns of linguistic diffusion”, Language Variation and Change, 5, 359-90. Behares, L.E. (1984), “Diglosia en la sociedad escolar de la frontera uruguaya con Brasil: matriz social del bilingüismo”, Cadernos de Estudos Lingüísticos 6, 228-34. ACTAS DO I SIMPOSIO INTERNACIONAL SOBRE O BILINGÜISMO 651

Bisol, L. (1992), “Palatalization and its variable restriction”, Journal of the Sociology of Language 89, 107-24. Bortoni-Ricardo, S.M. (1985), The urbanization of rural dialect speakers, Cambridge, Cambridge University Press. Elizaincín, A. (1978), “Bilingüismo y problemas educativos en la zona fronteriza uruguayo- brasileña”, Lingüística y Educación. Actas del IV Congreso Internacional de la Asociación de Lingüística y Filología de América Latina, 301-10. __ (1979), Precisiones sobre los dialectos portugueses del Uruguay, Montevideo, Universidad de la República. __ (1992), Dialectos en Contacto. Español y Portugués en España y América, Montevideo, Arca. Elizaincín A., L. Behares, G. Barrios (1987), Nós Falemo Brasilero. Dialectos Portugueses en Uruguay, Montevideo, Amesur. Head, B. (1981), “Social factors in the perception of phonetics differences”, Cadernos de Estudos Lingüísticos 2, 158-66. Hensey, F. (1972), The sociolinguistics of the Uruguayan-Brazilian border, The Hague, Mouton. __ (1982), “Spanish, Portuguese, and Fronteiriço: Languages in contact in North Uruguay”, International Journal of the Sociology of Language 34, 7-23. Herzog, M. (1965), The language in Northern Poland: Its geography and history, Bloomington, Indiana University Press. Labov, W. (1966), The social stratification of English in New York, Washington DC, Center for Applied Linguistics. __ (1972), Sociolinguistics patterns, Philadelphia, University of Pennsylvania Press. Le Page, R. (1980), “Projection, focussing and diffusion”, York Papers in Linguistics 9, 9-31 (first published in 1975). Rona, P. (1965), El dialecto fronterizo del Norte del Uruguay, Montevideo, UROU, Facultad de Humanidades y Ciencias. Thomason, S.G., T. Kaufman (1991), Language contact, creolization and genetic linguistics, Berkeley, University of California Press.